Questions about PS2

I think everyone can agree on the Saturn being a pain in the ass, same for the PS2, but after that the order is very debatable.
 
I think everyone can agree on the Saturn being a pain in the ass, same for the PS2, but after that the order is very debatable.
In some topics here people said than N64 was realy hard to program. Someone even said that after N64 VU on PS2 was just trivial.
 
In some topics here people said than N64 was realy hard to program. Someone even said that after N64 VU on PS2 was just trivial.

The MIPS CPU on the N64 was very easy to use. The hard part would be coding the RSP. But, the vast majority of games just used Nintendo's supplied "Fast3D" RSP microcode and didn't bother peeking inside themselves.

https://en.wikipedia.org/wiki/Nintendo_64_programming_characteristics#Microcode

http://level42.ca/projects/ultra64/Documentation/man/pro-man/pro25/index25.1.html

The Factor5 guys were really good at utilizing the RSP. There's also the lost story of the hero programmer who used the RSP to squeeze 2 CDs of FMV into a 64 MB cartridge for Resident Evil 2 for the N64. Unfortunately, both the source article and the only context article I can find are down. So, here are cached versions instead. Context. and Gamasutra article.
 
Last edited:
I think everyone can agree on the Saturn being a pain in the ass, same for the PS2, but after that the order is very debatable.

Videogame console systems have been making the impossible possible in terms of coding.

It's really tiring to read branded misquoted memes from 1995 on Saturn "game development"

The Saturn as it was and it's price only really failed due bad marketing moves and lack of marketing campaign moves.

The hardware development facts being spun into fear uncertainty and doubt by print magazines and later using a misquote on Yu Suzuki into a meme did a ton of damage but not as much as the lack of marketing Sega could have done in North America region.

Cancelling 32X and focus that funding into Saturn as a Cartridge and CD-ROM based console would also have helped big time as 32X only became a loss to tally.

Recently Digital Foundry went retro to analyze the effort of Lobotomy Software which made Power Slave, Quake Saturn and Duke Nukem Saturn which were proof that even if some still regurgitate the "nightmare to program for" there's gamers and tech fans justified in using the term "lazy devs"

The bad move by Sega was not funding L.S. or Core which made Tomb Raider...

Reading retrospectives claiming that Sonic was needed are overblown hyperbole.

Game programming school wasn't easy school...and Saturn games releases only multiplied and improved.

PS2 was definitely hard and difficult to program for as was N64.

Comparatively speaking the PS-X was really just well documented specially because Sony didn't have to worry about splitting efforts on 32X and Saturn and Model 2 Arcade development.

Once all the documentation is as complete as possible there is no excuse for dev teams to start cranking out amazing looking visuals and gameplay mechanics.

It's interesting to note that reading about the Sega 4MB ram Cartridge and N64 4MB expansion pack that both Sega and Nintendo didn't believe that devs should "require" the use to build 2d/3d engine games because they believed it would alienate consumers who "would not have been able to buy or be interested"

What's even more disturbing is that these "beliefs and decisions" were being made very early meaning that if those upgrades were instead released with freedom for ambitious and skilled devs who don't use premade level editors then call them games then the software that could have used those old expansion ram packs would have still become relevant in demonstrating the full potential of those ancient console hardware.

Saturn plus 4MB ram Cartridge plus experienced skilled devs would have resulted in the highest resolutions, locked 60fps best textures etcetera on both 2d and 3d polygon graphics based games becoming more numerous and that would have caused gaming consumers to notice and buy...specially in Japan where Saturn had a healthy install base and dedicated 3rd party relations.

N64 plus 4MB expansion pack plus skilled devs ignoring the ancient documentations and Nintendo being more open would also have resulted in the highest resolutions possible plus framerates and more so when cartridge sizes reached the biggest amounts possible...even the 64DD would have been able to be properly released if not for bad decisions on "oh that looks hard...we don't want devs to have a hard time..."

Of course prices on Rom chips falling and enabling higher cart sizes was unforseen...

Surely Turok sequels or even starting from Seeds of Evil would have jumped the visuals, surely a skilled Pilotwings 64 sequel and Starfox 64 sequel and Super Mario 64 sequel if made under such guidelines would have only added to sales and market domination even in light of PS-X domination back then.

Once a dev team is highly skilled and experienced (it helps too if they are first/second party) even the complex Saturn becomes trivial...because they already know the documentation.

Dreamcast was a total failure in that regard because it was just a blatant copy of PlayStation and Nintendo 64 with fondly remembered PlayStation hardware developed game as it's top fighting game.
 
Last edited:
In some topics here people said than N64 was realy hard to program. Someone even said that after N64 VU on PS2 was just trivial.
Depends what you're trying to do. Console A might be very easy to develop the easy stuff (time to first triangle) but put up walls when you try to do something a little outside its box, while Console B might be really difficult to use on the basics but no worse on the fancy stuff.
 
There was a late 90s maybe early 2000s (months before the PlayStation 2 launched) Next Generation or Edge print magazine cover story and article detailing that the Emotion Engine was very complex and hard to program for, perhaps the overall architecture as well.

There were many early games that lacked some features like the difference between the Japanese and North America region version of Tekken Tag Tournament which itself was a conversion (not a port) of the enhanced Namco System 12 Arcade board which was essentially an enhanced PlayStation 1 based board.

I believe often times the "easy to develop versus hard to develop for" gets caught up in the print magazine articles which were mostly filler opinion pieces in attempts to explain highly technical stuff...kinda like blog sites now.

I also have this feeling that because of all the Sega Dural/BlackBelt/Katana and 3dFX document leaks in late 1996-97 this actually caused Sony's hardware designs to be accelerated (I'd like to say rushed but it's incorrect) to meet the competition...plus a lot of the vaporware threats back then.

Hence that PS2 prototype engineering board photo that showed up in early 1999 print magazines was in development for a while. As well as Sony not really or not that they would have the belief that PS2 was gonna be such a success because even they didn't know it... the phenomenon of how PS2 became such a huge sales success was a major factor in turning that "complex, hard to dev for" platform into a fully funded (because there's an install base) well documented dev tools that helped Kojima Productions go from MGS2 to MGS3 which falls under what you said Shifty in that the game went beyond what the console was documented in capabilities.

I always wondered if Sega had been more secretive on Dural/Katana (perhaps never doing that 3dFX contract) and maintained proper support on Saturn then the Dreamcast would have been at least a bit more of a surprise...and perhaps Sony would have opted for a higher spec target with more or mainly double the VRam with PS2.

That would have probably made initial game development appear easier and later game development to be mind blowing but it'd have been a 2d monster, and higher textures.

But again console game development works with what finite resources are.
 
London-boy, Corysama, Akumajou, Shifty Geezer, thank you for answers and interesting talk. Do you think hard programing is more good or more bad in general? In my opinion consoles hard to program is better, because there is better results on those consoles in the end of lifecycle. First games is weird, low framerate etc. Next games is a little bit better, but still not good in terms of graphics. Then next games make huge jump. You see a lot more polygons, a lot better textures, great lighting, great post-process effects, stabi famerate etc. And last games is so good, what you can't understand how all this is possible on that console. For easy to program consoles situation is different. As an example first Xbox. First games already showed a lot, then there was some years when there were very litle graphics improvements in most games. And onle some last games (Half-Life 2, Doom 3, Far Cry Instincts), showed hage difference. So you got Xbox and it already have good graphics, which isn't bad, but in next years console can't surprise you somehow.
Same situation with PS3 and PS4. First games on PS3 was very weird, and then, in next year... Uncharted... WOW! Next year more great graphics, then Killzone 2, God of War 3, Uncharted 2! And it's just incredible! And in the end of lifecycle Uncharted 3, God of War Ascension, The Last of Us. And you just can't understand how it's possible what there was some probems in first games, when you see those last games. PS4, first games already great (Killzone Shadowfall, Driveclub), and in more than two years there is only little improvemets. Of course Ratchet & Clank is great, and Uncharted 4 is great, but there is no such effect like on hard to programm consoles.
Yes I know what developers prefer easy to program consoles, but isn't that good also to have some challenge? :D
 
I strongly disagree that a platform that is hard to program for is in any way 'better' as it supposedly gives the impression that later in the the lifecycle there is an obvious progression in the quality of the games.

A platform should be able to churn out great looking/sounding games without making life difficult, and with the minimum effort necessary.

Prime example I can think of is Killzone Shadow Fall. Launch game and it was just beautiful. No messing around with the whole 'it will look so much better in 4 years!' - which can still happen, of course, as there is always a learning curve with any platform.

Game developers need to spend time making a game, creating beautiful assets, stories and gameplay. Not fighting with the hardware trying to get their game to work properly.
 
Prime example I can think of is Killzone Shadow Fall. Launch game and it was just beautiful. No messing around with the whole 'it will look so much better in 4 years!' - which can still happen, of course, as there is always a learning curve with any platform.
Yes, but as I said PS4 was WOW already at start, but it can't really impress me now. PS2 and PS3 did it every year.

Game developers need to spend time making a game, creating beautiful assets, stories and gameplay. Not fighting with the hardware trying to get their game to work properly.
But they mostly don't. There isn't so much great and impressive games now. Hard platform makes all work harder and in he end there's better results. But it's my opinion.
 
The Factor5 guys were really good at utilizing the RSP. There's also the lost story of the hero programmer who used the RSP to squeeze 2 CDs of FMV into a 64 MB cartridge for Resident Evil 2 for the N64. Unfortunately, both the source article and the only context article I can find are down. So, here are cached versions instead. Context. and Gamasutra article.
Absolutely great achievement! Very impressive!
 
Yes, but as I said PS4 was WOW already at start, but it can't really impress me now. PS2 and PS3 did it every year.


But they mostly don't. There isn't so much great and impressive games now. Hard platform makes all work harder and in he end there's better results. But it's my opinion.
What games have you played? Sorry but a few games released this year are extremely impressive.

This whole "games are 20 times better at the end of the lifecycle" is strictly related to the whole "secret sauce" movement, which really needed to die a few years ago. And now with 3-4 year long generations, I'm sorry to say you'll never see again. Thank god.
 
Even on platforms that are "easy to develop for", you see changes year by year. Not always in the sense of strict improvements, but methodologies get refined and expanded, compromises and priorities rethought. The changes from year to year remain interesting, if in some respects not as immediately striking, and even if you get some instances where some people feel that the compromization made for an overall downgrade. Heck, that last bit is part of what makes it interesting.
 
Personally I'd rather have 7+ years of Awesome Games capable of fully using the hardware rather than 2 years of Awesome Games and the rest various shades from okay to poor because devs are wrestling with using the thing. And I can't see why anyone would purposefully want their experience to be hampered by difficulty in the first years either. If that's really what you want, make the machine easy to develop for and put a software cap on exposing only 50% in the first year, upped 10% each year afterwards so everyone can marvel at how the games get better over time. This has the same result without stressing out developers.
 
Personally I'd rather have 7+ years of Awesome Games capable of fully using the hardware rather than 2 years of Awesome Games and the rest various shades from okay to poor because devs are wrestling with using the thing. And I can't see why anyone would purposefully want their experience to be hampered by difficulty in the first years either. If that's really what you want, make the machine easy to develop for and put a software cap on exposing only 50% in the first year, upped 10% each year afterwards so everyone can marvel at how the games get better over time. This has the same result without stressing out developers.
The trick is to time cap removals right before the releases of AAA games.

"We had frequent dips to 25fps, but Sony just unlocked the CPU's multiply instructions, and now we're holding a rock-solid 30! Wow!"
 
London-boy, Corysama, Akumajou, Shifty Geezer, thank you for answers and interesting talk. Do you think hard programing is more good or more bad in general?

Short answer: No. Easier plz.

Long Answer: "Making it harder gets better results." is something I hear from a lot of people. I think there are a few things that contribute to this trope.

1) A hard barrier to entry is a filter that only let through people who are predisposed to investing heavily. This seems to increase quality. But, in reality the hard-core would have done the same or better if it was easier. It's just that it leads to a lot less low-investment work being done additionally. So, peak quality is lower, but it feels better because there is less noise and median quality is higher.

2) Difficult systems are better utilized over time. This is a bit of survivorship bias. The Saturn and Jaguar were both very difficult. I don't think anyone would say their games were better for it. The PS2 and PS3 being difficult meant that the early games were crappier than they should have been. But, market forces made it worthwhile to continue investing anyway. So, late in the generations the games showed what should have been possible all along.

3) Difficult systems are often difficult because they are trying to make difficult problems possible to solve. This is the only non-fallacy of the list. If everything is easy, then you probably could have attempted something better that was harder. For a given amount of designer brilliance, there will be tradeoffs between making common things easier vs making harder things practical. The PS3's SPUs were an to make a lot of difficult problems practical in the face of memory and CPU engineering realities. Unfortunately, in practice most of the SPU work seemed to go towards helping the PS3's GPU to keep up with the 360's GPU.
 
PS2 was definitely hard and difficult to program for as was N64.

How could you contend that Saturn wasn't hard to program for and these was just media FUD while at the same time saying PS2 and N64 were legitimately challenging?

Saturn was extremely complex and compared to PS1 you had to do a lot more work to get 3D out of it. It wasn't just a matter of documentation, I'd say both consoles had lackluster documentation.

On PS1 the GTE did most of the transform, lighting, and projection work and gave you good feedback for depth sorting and clipping. Saturn had no such thing, you had to do all of that in software on one of the SH2s, which didn't even have proper support for division. The SCU DSP may have been designed with that purpose in mind but it didn't really work for it and few games used it for anything at all. End result is that most Saturn games don't even have lighting.

Quads were a bad fit for a lot of geometry and because they were actually distorted sprites (so axis aligned rectangular textures) they couldn't be clipped or tessellated very well. The guard band was more limited than PS1's so lack of clipping was a hindrance, not to mention that you still paid a significant cost for pixels that were in part of the guard band. And then there's the matter of the VDP1's broken blending and the kind of hoops you had to work through to try to get 3D blended on the VDP2 instead.

Trying to split code to run on two homogeneous CPUs was a big challenge at the time, especially since unlike in proper SMP systems they weren't cache coherent and there weren't atomic memory primitives. The caches were small, which would usually be okay when the memory wasn't that comparatively slow, but with two processors that made RAM contention a real issue.

PS1 also had a DCT decoder that Saturn lacked, so they had to decode in software and quality tended to suffer. And things like Saturn's 2MB of RAM being split into separate 1MB address regions were little headaches.
 
What exactly was the cause of the texture warping as quads went off-screen?

Partially off-viewport sprites weren't a problem in 2D games, so what was it about 3D (distorted quads) that caused issues? Was the guard band too small for very large quads, or too slow...?
 
Personally I'd rather have 7+ years of Awesome Games capable of fully using the hardware rather than 2 years of Awesome Games and the rest various shades from okay to poor because devs are wrestling with using the thing.
But don't you think that consumers will say "hey, we need new console, because on that one graphics doesn't change for some years? :D

But, market forces made it worthwhile to continue investing anyway. So, late in the generations the games showed what should have been possible all along.
But as I said before, if all that is avaiable from start, people could say, what they need new console, or say that devs are lazy and don't use all power of machine.

Saturn was extremely complex and compared to PS1 you had to do a lot more work to get 3D out of it. It wasn't just a matter of documentation, I'd say both consoles had lackluster documentation.
Great info about Sega Saturn, thanks!
 
But don't you think that consumers will say "hey, we need new console, because on that one graphics doesn't change for some years? :D

Consoles are a precarious mix of developers becoming more experienced with extracting performance from the hardware fighting against increasing demand of games wanting to do more. With some exceptions, late console generation games generally look way better than early console generation games: GTA III to GTA San Andreas, Uncharted to The Last of Us, GTA IV to GTA V.

Even if you have a machine with a basic architecture and subsystems that interact very little, which you can learn and master really quickly, software techniques and algorithms are always progressing.
 
Back
Top